On-line evaluation of equalizer performance based on high-order statistics

In this paper, a simple recursive method is developed in order to estimate the bit-error-rate (BER) at the output of an adaptive equalizer. This is accomplished in two steps. Firstly, a theoretical procedure for detecting decision errors is derived. This test is then used with an adaptive blind channel identification scheme, based on high-order statistics theory, in order to derive the final adaptive BER estimator. Simulation results point out that our method provides on-line and reliable BER estimation with a low computational burden for linear channels models.